Revised December 4, 2001

Attractions of Payroll Tax Holiday Diminish Upon Closer Examination
by Robert Greenstein and Peter Orszag(1)

PDF of report

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

Interest has been increasing on Capitol Hill in a proposal to suspend the Social Security payroll tax for one month. This proposal is being considered for possible inclusion in a stimulus package.

At first blush, this proposal may sound quite attractive. Upon closer examination, however, the proposal turns out to be problematic. Although costly, it is likely to be relatively ineffective as a stimulus measure.



To be effective, stimulus should be delivered sooner rather than later. Although the economic outlook remains uncertain, private-sector forecasts suggest the economy is likely to be recovering by the middle of next year.

According to a recent analysis issued by several associations of payroll managers, however, a tax holiday "would require 6 months or more to ensure successful implementation."(2) According to the analysis, payroll processors and others involved in payroll preparation are particularly busy near the end of a calendar year and in the month of January in testing and implementing changes for the upcoming year, as well as in preparing and submitting employment tax and wage reports. This makes implementation of the proposed holiday particularly difficult in these months.

For this and other reasons relating to complexities involved in implementing a payroll tax holiday, it appears unlikely a holiday could be successfully implemented in January, as some Members of Congress appear to be assuming. It could well be a number of months until the holiday could be put into effect. The longer the delay, the less the value of the measure as a stimulus.


Inefficient Business Tax Break

The proposal would provide a holiday from both the employer and the employee shares of the payroll tax, each of which amounts to 6.2 percent of eligible wages. The holiday on the employee share will immediately translate into higher take-home pay, and thus provide a cash infusion to families. But the holiday from the employer share of the payroll tax will not readily translate into higher take-home pay and will be largely ineffective as a stimulus measure.

Standard economic analysis suggests that over the long run, a permanent reduction in the employer payroll tax would increase wages. (The reason is that firms care about the total cost of labor, not how much comes from payroll taxes and how much from wages. If payroll taxes are reduced on a permanent basis, wages should eventually increase in line with the payroll tax reduction, all else being equal.) But such adjustments in wages would take time to occur. It is extremely unlikely that a month-long holiday on the employer share of the payroll tax would immediately and fully translate into higher wages. To the contrary, most, if not all, of the holiday on the employer share of the payroll tax would simply represent an infusion of cash into businesses.

Infusing cash into businesses in this manner would not constitute effective stimulus. The primary problem that U.S. firms now face is a reduction in demand for their products, not a shortage of cash. Indeed, from the end of 1999 to the middle of 2001, holdings of liquid financial assets by nonfarm, nonfinancial corporations rose by more than $100 billion (or more than 17 percent), and holdings of total financial assets by such firms rose by more than $700 billion (or almost 9 percent).(3) Another business tax cut included in most stimulus packages — a tax break for new investments that businesses make — would be stimulative because it is directly tied to, and provides an incentive for, new investment. But indiscriminately injecting cash into firms through a tax break without linking the tax break to new business activity, as a one-month payroll tax holiday would do, would not have much effect in stimulating the economy.(4) Such tax breaks would instead tend to be retained by businesses or fed through into additional cash for corporate shareholders and business owners, who tend to be disproportionately higher-income. Since higher-income households tend to spend relatively little of any additional income they receive, not much additional demand for goods and services would be created

About half of the estimated $38 billion cost of the proposal — the employer share — thus would represent an indiscriminate infusion of cash into businesses and be relatively ineffective as a stimulus.(5) This half of the proposed reduction in the payroll tax would largely result in a windfall for businesses without a commensurate near-term benefit for the economy.


The Employee Side of the Equation

Even the employee side of the payroll tax reduction would not be particularly stimulative. Under the payroll tax holiday proposal that Senator Pete Domenici and Senate Republican leaders have advanced, the payroll holiday would replace the tax rebate for low-income workers that has been part of all major stimulus proposals until now, including the Senate Finance Committee legislation, the House bill, and the Bush Administration plan. Yet the low-income rebate would be much more progressive than the payroll holiday — and thus more efficient as a stimulus measure — for two reasons: It would cover only low-income workers, and it would provide the same dollar amount to all low-income workers it assisted.(6) By contrast, the payroll holiday also would cover workers at higher income levels and would provide larger dollar cuts to those workers than to those with lower incomes.

For example, a worker earning $10,000 a year would receive a little more than $50 from the one-month payroll holiday, as compared to $300 from the low-income rebate. A worker earning $80,000 a year, by contrast, would receive more than $400 from the payroll holiday, as compared to nothing under the low-income rebate. The payroll holiday is substantially less progressive than the low-income rebate.

The progressivity of the tax cut matters because it is closely connected to the measure's stimulative impact. Lower-income households tend to spend a higher percentage of any additional income they receive than do high-income households. To maximize the bang for the buck in bolstering the economy in the short run, a tax cut for households should encourage people to spend the new funds the tax cut provides, rather than save these funds. As a result, the more focused any tax cut is on lower-income households, the better. Since the payroll tax holiday would not be particularly progressive, it would not be particularly effective at stimulating the economy in the short run.

The first round of rebate checks were distributed in a more progressive manner than a payroll holiday would be: For those with sufficient income to qualify for the full rebate, the rebate amount did not depend on income. Therefore, the rebate directed relatively more of its dollars toward moderate earners than would the proposed payroll tax reduction, which would be based on a percentage of an individual's earnings, up to the maximum Social Security earnings level. Nonetheless, a large portion of the rebates appears to have been saved. For example, a survey undertaken by the University of Michigan in September and August found that only 19 percent of respondents said that they were going to spend their rebates, with the remainder either planning to pay off debts or save the funds.(7) The personal saving rate also rose significantly in July, August, and September, which is consistent with a large portion of the rebates being saved rather than spent.(8)


Squeezing Out More Effective Stimulus Items

There is a considerable risk that if a payroll tax holiday is adopted, assistance for the unemployed would be squeezed. Consider the math. The payroll tax holiday would cost $38 billion. Negotiators on the stimulus package are likely to agree to let businesses immediately write off (or "expense") 20 percent or 30 percent of the cost of new investments. That will cost $26 billion to $39 billion in fiscal year 2002. The result is a combined cost of $64 to $77 billion in fiscal year 2002. And that figure does not include the cost of a number of smaller tax-cut provisions that also are on the table, from changes in the corporate Alternative Minimum Tax, to loss carry-back provisions, other depreciation provisions, and tax extenders. Yet there is apparent agreement that the cost of the overall package in fiscal year 2002 should be held below $100 billion. Including the payroll tax holiday (at a cost of $38 billion) could thus result in fewer resources being devoted to improvements in help for the unemployed, such as expansions of unemployment insurance benefits. Improvements in unemployment benefits not only would be better targeted on those in need than a payroll tax holiday, but also would constitute more effective and efficient stimulus.(9)

Would High-Income People Benefit from the Proposed Tax Holiday?

Senator Domenici's original payroll tax plan called for the holiday to occur in December. Since higher-income people reach the Social Security payroll tax cap and stop paying Social Security payroll taxes before December, they would not benefit.

Now that the proposal would presumably not be implemented until some time after December, the effect of the proposal on high-income individuals is less clear. It is technically possible to devise the payroll tax holiday so that higher-income people (who would be likely to save most of their payroll tax cut rather than spend it) do not secure significant payroll tax reductions. That, however, would make administration of the payroll tax holiday more complex and could further delay its implementation.



A payroll tax holiday may initially seem like a sound idea as a stimulus proposal. Payroll experts suggest, however, that it could not be implemented in a timely fashion. Furthermore, roughly half of the cost of the provision would represent a relatively ineffective infusion of cash into businesses, while the other half would be substantially less effective as a stimulus than the low-income rebate that was earlier endorsed by the House of Representatives, the Senate Finance Committee, and the White House.

End Notes:

1.  Peter Orszag is the Joseph A. Pechman Senior Fellow in Tax and Fiscal Policy at the Brookings Institution.

2. The American Payroll Association, the American Society of Payroll Management, the National Payroll Reporting Consortium, and the Society for Human Resource Management, "Implementation Problems Related to a 'Payroll Tax Holiday' in December 2001 or January 2002," November 2001.

3. Federal Reserve Board, "Balance Sheet of Nonfinancial Corporate Business," Table B.102, September 18, 2001.

4. See Peter Orszag and Joseph Stiglitz, "Tax Cuts Are Not Automatically the Best Stimulus: A Response to Glenn Hubbard," Center on Budget and Policy Priorities, November 27, 2001. The temporary drop in the total cost of labor that firms would incur as a result of the payroll tax holiday could encourage some firms to hire more workers, but it seems unlikely many firms would respond in a significant manner to a one-month decline in the cost of labor.

5. It should be noted that the business tax cut which a payroll tax holiday would provide would likely be more stimulative than repeal of the corporate Alternative Minimum Tax. The proportion of the business tax cut that would go to small firms would be greater under a payroll tax holiday than under corporate AMT repeal. To the degree that any cash constraints are affecting business activity, such constraints are more likely to occur among smaller firms than larger ones.

6. It also would provide a partial rebate to moderate-income workers who received less than $300 under the initial rebates distributed this fall.

7. Matthew D. Shapiro and Joel Slemrod, 2001, " Consumer Response to Tax Rebates," mimeo, October.

8. Personal saving averaged 1.0 percent of disposable personal income in 2000. In June 2001, the personal saving rate also was 1.0 percent. It then rose to 2.4 percent in July, when the rebate checks began to be mailed, 4.2 percent in August, and 4.6 percent in September. After the mailing of the rebate checks largely ended, the personal saving rate dropped substantially in October. See Bureau of Economic Analysis, "Personal Income and Outlays: October 2001," December 3, 2001, Table 2. Although other factors may explain some of the increase in the personal saving rate, it seems likely that the rebates played a role.

9. Peter Orszag, "Unemployment Insurance as Economic Stimulus," Center on Budget and Policy Priorities, November 15, 2001.